Efficient Multijunction Solar Cell Design for Maximum Annual Energy Yield by Representative Spectrum Selection

Abstract

We describe a systematic approach to multijunction solar cell (MJSC) design that unambiguously identifies the spectrum to be used in cell optimization such that local annual energy yield is maximized. A set of candidate spectra is generated from air mass (AM) values ranging from AM1d to AM5d. Each candidate spectrum is used to find the bandgap combination that maximizes cell efficiency and its energy yield is then calculated using an efficient data reduction approach. The bandgap combination that maximizes annual energy yield identifies the representative spectrum. We do this for cells with up to eight junctions across all clear-sky latitudes and compare our results to other cell optimization approaches. Our representative spectrum selection (RSS) approach is robust and highly tolerant of variations in latitude, particularly when compared to the standard AM1.5d approach which, at midlatitudes, cannot be used without suffering an increasingly severe yield penalty. Comparison against the 50% cumulative energy AM (50% AM) design approach is enabled by using the same design conditions (sea level and ASTM standard atmosphere) in order to unambiguously associate each 50% AM value with a cell design spectrum. We find that our RSS approach always produces cells with slightly higher annual energy yields than are achieved by the corresponding 50% AM designs. While both approaches show similar yields for devices with few junctions, we find yield enhancements approaching 1% for cell designs with many junctions, emphasizing the need to consider the spectral variability of the local solar resource. This consideration is systematically enabled by our RSS approach, addressing a deficiency in the previous design approaches.